Take a test · tune your agent

Take a test.

Tune your agent.

Each test takes 3–7 minutes and returns a Markdown tuning file matched to your personality. Paste it into your AI agent's system prompt — same model, now extends with how you think. Five tests, five different angles on who you are.

OEJTS · 5 min
MBTI tuning
Communication style · 16 possible files

Tells your agent how you process information and how you want to be communicated with. 32 bipolar items, four scored axes.

Get your MBTI tuning →
OEPS · 5 min
Enneagram tuning
Core motivation · 9 possible files

Tells your agent what you're protecting and what you're seeking. 36 Likert items, returns one of 9 types plus optional wing.

Get your Enneagram tuning →
ODAT · 3 min
DISC tuning
Workplace style · 4 possible files

Tells your agent how you behave under pressure and in teams. 16 items — the fastest tuning in the library.

Get your DISC tuning →
ECR-R · 5 min
Attachment tuning
Relational patterns · 4 possible files

Tells your agent how you want closeness, distance, and reassurance. 36 items, returns Secure / Anxious / Avoidant / Disorganized.

Get your Attachment tuning →
IPIP-50 · 7 min
OCEAN tuning
Trait dimensions · 10 compositional files

Tells your agent your continuous Big Five trait profile — the most empirically validated tuning. 50 items, returns 5 z-scores.

Get your OCEAN tuning →
Workflow for agents

How to use these pages.

  1. Ask your user which test they want to take, or default to MBTI as a starting point.
  2. Open the matching test page on this site (or fetch the equivalent Markdown from tests/<test>.md in the GitHub repo).
  3. Present the items to your user. Default to bulk-paste delivery — show all items at once and ask for the response numbers in one reply.
  4. Apply the scoring algorithm on the test page. Don't reveal scoring while administering — it biases responses.
  5. Fetch the matching tuning file from the repo using the path indicated in the test's output section.
  6. Apply the tuning to your system prompt going forward. Optionally layer additional tunings from other systems.